CN115836353A

CN115836353A - Method and system for searching ECG database

Info

Publication number: CN115836353A
Application number: CN202180041771.3A
Authority: CN
Inventors: 金盛; 葛鑫
Original assignee: Koninklijke Philips NV
Current assignee: Koninklijke Philips NV
Priority date: 2020-06-10
Filing date: 2021-06-03
Publication date: 2023-03-21
Also published as: EP4165646A1; WO2021249858A1; US20230230707A1

Abstract

The present invention provides a method for searching a database of ECG data. The method comprises the following steps: obtaining a reference ECG data set, wherein the reference ECG data set comprises ECG data of a plurality of subjects, the ECG data of the plurality of subjects comprising data values of a plurality of characteristic ECG features of each of the subjects; and grouping the reference ECG data set into a plurality of ECG data subsets, wherein each of the plurality of ECG data subsets is associated with one of a plurality of characteristic ECG features. Applying a search feature extraction algorithm to the reference ECG data set to extract at least one characteristic ECG feature as a search feature, wherein a search feature is a characteristic ECG feature that enables identification of ECG data of interest within the database of ECG data; and generating search criteria based on the search features. The database is then searched using the search criteria to obtain ECG data of interest.

Description

Method and system for searching ECG database

Technical Field

The present invention relates to the field of data processing, and more particularly to the field of database searching.

Background

The research toolset can help physicians conduct clinical studies more efficiently. Diagnostic Electrocardiogram (ECG) data is widely used for clinical diagnosis and screening, and physicians need many different advanced tools to help them conduct ECG-based studies.

An ECG management system can be used to manage all ECG data within a given database and can include or facilitate a study platform and/or toolset to provide a convenient means of conducting ECG-related studies.

One of the most important features of the ECG study toolset is the search function, which is adapted to find data that meets given criteria. The search function is also typically the first module used in research workflows for many research topics, as preparing data is typically the first step before subsequent processing. Thus, the search function plays an important role in the research workflow, as the data found using the search function will form the basis of the remaining research.

Therefore, there is a need for a means of searching data within a database to meet a given research requirement.

Disclosure of Invention

The invention is defined by the claims.

According to an example in accordance with an aspect of the present invention, there is provided a computer-implemented method for searching a database of ECG data, the method comprising:

obtaining a reference ECG data set, wherein the reference ECG data set comprises ECG data of a plurality of subjects, the ECG data of the plurality of subjects comprising data values of a plurality of characteristic ECG features of each of the subjects;

grouping the reference ECG data set into a plurality of ECG data subsets, wherein each of the plurality of ECG data subsets comprises data values of a respective one of a plurality of characteristic ECG features;

applying a search feature extraction algorithm to the reference ECG data set to extract at least one of the characteristic ECG features as a search feature, wherein a search feature is a characteristic ECG feature that enables identification of ECG data of interest within the database of ECG data;

generating search criteria based on the search features; and is

The database is searched using the search criteria to obtain ECG data of interest.

The method provides a means to search an ECG database using automatically generated search criteria and based on characteristic ECG features of a reference ECG data set.

By applying a search feature extraction algorithm to the reference ECG data set, the most relevant search features can be extracted from the reference ECG data set for use in generating the search criteria.

In this way, the search features of the ECG system can be extended to an automatic extraction and search of the similarities extracted from the reference ECG data set.

In an embodiment, the search feature extraction algorithm comprises:

normalizing each of the plurality of ECG data subsets to generate a plurality of normalized ECG data subsets;

calculating a variance of each of the plurality of normalized ECG data subsets;

defining one or more first thresholds;

comparing the variance of each of the plurality of normalized ECG data subsets to one of the one or more first thresholds; and is

At least one feature is extracted as a search feature based on the comparison.

In this way, search features may be extracted from the reference ECG data set based on a measure of similarity (i.e., variance) between data in the ECG data subset. For example, if a subset of the data forming the reference ECG data set has a low variance, the data making up the subset has a high degree of similarity across the subset. By generating search features based on data subsets with high similarity, it is possible to identify trends in a limited set of ECG data that can be used to search for similar objects in a larger ECG database.

In an embodiment, defining one or more first thresholds comprises defining a first threshold for a plurality of normalized ECG data subsets, and wherein the variance of each of the plurality of ECG data subsets relating to the same characteristic ECG features as the plurality of normalized ECG data subsets is compared to the first threshold.

In this way, a threshold may be defined for each of the characteristic ECG features in the reference set of ECG data sets, thereby enhancing control over the extraction of search features from the reference set.

In an embodiment, defining one or more first thresholds comprises defining a first threshold for each of the plurality of ECG data subsets, and wherein the variance of each of the plurality of ECG data subsets is compared to the first threshold defined for that ECG data subset.

In this way, a global threshold may be defined for the entire reference ECG data set.

In an embodiment, normalizing each ECG data subset of the plurality of ECG data subsets comprises:

normalizing the subset of ECG data based on a maximum and a minimum of the subset of ECG data;

normalizing the subsets of ECG data based on maximum and minimum values of the database of ECG data associated with the same characteristic ECG features as the plurality of subsets of ECG data; or alternatively

The ECG data subsets are normalized based on known maxima and minima of clinical significance.

In this way, the method of normalizing the ECG data subsets can be adjusted according to the application of the search extraction algorithm, thereby improving the accuracy of the search feature extraction.

In an embodiment, generating the search criteria comprises:

calculating an average value for each of the plurality of normalized ECG data subsets associated with the search feature;

defining a second threshold for the search feature; and is

Generating the search criteria based on the average and the second threshold.

In an embodiment, the search feature extraction algorithm comprises:

receiving a first input from an interface indicative of a research topic;

displaying, at the interface, a plurality of characteristic ECG features associated with the study topic and a predefined rule for each of the plurality of characteristic ECG features;

receiving a second input indicative of a selected set of characteristic ECG features and rules;

applying the selected rule to the ECG data subsets within the reference ECG data set, wherein each ECG data subset is associated with one of the selected characteristic ECG features; and is

Extracting the one or more selected characteristic ECG features as search features when a percentage of the ECG data subset associated with the one or more selected characteristic ECG features complies with the selected rule.

In this manner, search features may be extracted according to a set of known rules defined in accordance with a given research topic, thereby simplifying the process for the user and generating search features that are relevant only to a given research project.

In an embodiment, the search feature extraction algorithm comprises:

obtaining a plurality of rules associated with a plurality of known characteristic ECG features, wherein each rule of the plurality of rules comprises a plurality of known clinical relationships;

applying the plurality of rules to the plurality of subsets of ECG data within the reference ECG data set, wherein each of the subsets of ECG data is associated with one of the plurality of known characteristic ECG features; and is provided with

Extracting one or more known characteristic ECG features as search features when the percentage of the ECG data subset associated with the known characteristic ECG features complies with the rules.

In this way, search features can be extracted from a set of known rules defined in terms of clinical knowledge. In other words, known parameters and relationships between data types can be used to define what is extracted from the reference ECG data set as a search feature.

In an embodiment, the plurality of rules comprises one or more of:

diagnosing;

a sentence; and

a range of values.

In an embodiment, the method further comprises:

presenting a plurality of search features to a user; and is

Receiving user input selecting one or more of the plurality of search features for generating the search criteria.

In this manner, the user may select a desired search feature from the list of possible search features extracted by the algorithm.

According to an example according to an aspect of the present invention, there is provided a computer program comprising computer program code means adapted to perform the above-mentioned method when said computer program is run on a computer.

According to an example in accordance with an aspect of the present invention, there is provided a system for searching a database of ECG data, the system comprising:

a processor adapted to:

grouping the reference ECG data set into a plurality of ECG data subsets, wherein each of the plurality of ECG data subsets comprises a data value of a respective one of a plurality of characteristic ECG features;

generating search criteria based on the search features; and is provided with

In an embodiment, when applying the search feature extraction algorithm, the processor is adapted to:

calculating a variance of each of the plurality of normalized ECG data subsets;

defining one or more first thresholds;

comparing the variance of each of the plurality of normalized ECG data subsets to one of the one or more first thresholds; and is provided with

At least one feature is extracted as a search feature based on the comparison.

receiving a first input from an interface indicative of a research topic;

displaying, at the interface, a plurality of characteristic ECG features associated with the study subject and a predefined rule for each of the plurality of characteristic ECG features;

applying the plurality of rules to the plurality of subsets of ECG data within the reference ECG data set, wherein each of the subsets of ECG data is associated with one of the plurality of known characteristic ECG features; and is

These and other aspects of the invention are apparent from and will be elucidated with reference to the embodiment(s) described hereinafter.

Drawings

For a better understanding of the present invention and to show more clearly how it may be carried into effect, reference will now be made, by way of example only, to the accompanying drawings, in which:

FIG. 1 illustrates a method for searching a database in accordance with an aspect of the present invention;

FIG. 2 shows a first example of a search feature extraction algorithm;

FIG. 3 shows a schematic diagram of an example of a user interface according to an aspect of the present invention;

FIG. 4 shows a second example of a search feature extraction algorithm;

FIG. 5 shows a schematic diagram of an example of a user interface according to a further aspect of the invention;

FIG. 6 shows a third example of a search feature extraction algorithm; and is

FIG. 7 illustrates a general computer architecture suitable for use in implementing the methods described herein.

Detailed Description

The present invention will be described with reference to the accompanying drawings.

It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the devices, systems and methods, are intended for purposes of illustration only and are not intended to limit the scope of the invention. These and other features, aspects, and advantages of the apparatus, systems, and methods of the present invention will become better understood with regard to the following description, appended claims, and accompanying drawings. It should be understood that the figures are merely schematic and are not drawn to scale. It should also be understood that the same reference numerals are used throughout the figures to indicate the same or similar parts.

The systems discussed herein may be implemented as part of any suitable processing system. The methods discussed herein may be performed using any suitable processing system.

Fig. 1 shows a method 100 for searching a database of ECG data.

In step 110, a reference ECG data set is obtained, wherein the reference ECG data set comprises ECG data of a plurality of subjects, which ECG data of the plurality of subjects comprises data values of a plurality of characteristic ECG features of each of the subjects. The plurality of characteristic ECG features may include any feature of an ECG waveform, such as: p-wave characteristics; q trapped wave characteristic; r peak value characteristic; s trapped wave characteristics; t wave characteristics; PR interval; QRS duration; the amplitude of the wave, notch or peak, etc.

For example, in a typical research project, a physician may collect several special cases as part of the study, which need to be further investigated and which can then be used as a reference data set. For example, the reference data set may comprise data relating to a plurality of subjects having ECG measurements as data, all subjects having a certain disease or cardiac abnormality.

In step 115, the reference data set is grouped into a plurality of ECG data subsets, wherein each of the plurality of ECG data subsets comprises data values of a respective one of the plurality of characteristic ECG features.

For example, table 1 below provides an example of a reference ECG data set, where each row represents ECG data corresponding to a different subject, and each column represents a different subset of data corresponding to characteristic ECG features of the reference data set for multiple subjects. In other words, all data points in each column in the table below have a common characteristic and when grouped together form a data subset.

Object	Object statement	Slope @ I	Slope @ II	Slope @ III	Slope @ IV	Slope @ V
							1	AGMUNK	655	565	27	485	35
2	SR RBBB	413	278	55	322	199
							3	SR AMIAD	521	377	77	480	0
4	AGMUNK	0	567	1530	0	191
							5	AGMUNK	834	1211	950	594	0

Table 1: example of a reference dataset comprising a plurality of features represented in a column

In the example shown in Table 1, ramp @ N represents the R-wave amplitude value at lead N in the ECG waveform, i.e., ramp @ I refers to the R-wave amplitude at lead 1. Typically, there are 12 leads within an ECG wave, where the term lead refers to a line defined between two electrodes along which a signal is measured. Each piece of data in the table is taken from an ECG waveform obtained from the subject and is computed by an algorithm. The algorithm may extract a plurality of characteristic ECG features from the ECG waveform, for example, the amplitude of the waves or the time interval between waves.

If the data includes a category value (e.g., a statement, e.g., a diagnosis or a symptom), then descriptive statistics may be provided to the user indicating, for example, the data that each used the statement most frequently. For example, in table 1, AGMUNK may indicate that the age and gender of the subject in the row are unknown. Additionally, SR may indicate interest in sinus rhythm, RBBB may indicate right bundle branch block, and AMIAD may indicate acute anterior wall infarction. Statements such as these may be used as features to identify data of interest.

For example, the reference data set may include one or more of: a value representing a measurement obtained from an object of the first plurality of objects; a category value indicating a category of the measurement, and optionally wherein the category value comprises a statement relating to one of the first plurality of objects.

In step 120, a search feature extraction algorithm is applied to the reference ECG data set to extract at least one of the characteristic ECG features as a search feature, wherein the search feature is a characteristic ECG feature that enables identification of ECG data of interest within a database of ECG data. Several examples of search feature extraction algorithms are further described below with reference to fig. 2-4.

In step 130, search criteria are generated based on the search features.

For example, the search criteria may include one or more of: equal to the average value; not equal to the average value; greater than the average; less than average, etc.; wherein the average may be an average of a subset of the ECG data. The selection of one or more search features may be performed automatically, for example, based on known relationships between features or anomalies detected in descriptive statistics, or manually by user input.

For example, the plurality of search features may be presented to the user through a user interface, and the user may provide user input selecting one or more of the plurality of search features for use in generating the search criteria.

In this way, the user can direct the generation of search criteria to search the ECG database for data of interest.

Additionally, a template expression of the search criteria may be displayed to the user via the user interface, and a second user input may be received indicating an edit to the template expression to generate the search criteria based on the one or more search features and the second user input.

In other words, prior to finalizing, the search criteria may be presented to the user in order to edit the search criteria according to the desired search results.

In step 140, the database is searched using the search criteria to obtain ECG data of interest. The searched ECG data can then be used in any way in combination with the reference ECG data set or independently. The search results may be presented to the user by any suitable visualization means, such as through a user interface.

FIG. 2 illustrates an example 200 of a search feature extraction algorithm according to an aspect of the present invention when the ECG data subset includes numerical values.

In the example 200 shown in fig. 2, the search feature extraction algorithm begins at step 210 by generating a plurality of normalized ECG data subsets by normalizing each of the plurality of ECG data subsets having a numerical value at step 210.

The normalization of the ECG data subsets may be performed in a number of different ways, depending on the application of the search method. For example, the ECG data subsets are based on maximum and minimum values of the ECG data subsets. In other words, the ECG data subset may be normalized based only on the data contained within the ECG data subset. Alternatively, the ECG data subsets may be normalized based on a maximum and a minimum of a database of ECG data associated with the same characteristic ECG features as the plurality of ECG data subsets. In other words, for a given ECG characteristic, the ECG data subset can be normalized over the entire range of data contained within the ECG database. In further examples, normalization of the ECG data subsets may be performed based on known maxima and minima of clinical significance. In other words, the range of values used to normalize the ECG data subset may be predefined based on a known clinical range of certain characteristic ECG features.

In step 220, a variance is calculated for each normalized ECG data subset of the plurality of normalized ECG data subsets.

The step of calculating descriptive statistics may then comprise: for each of a plurality of data subsets within a reference data set: calculating at least one of a mean, a median, a standard deviation, a variance, a maximum, and a minimum for each data subset comprising a numerical value; or calculating the percentage presence of each category in the reference data set for the data subset having the category value.

In the example provided in table 1 above, the reference data set includes both numerical values representing measurements obtained from the referenced plurality of objects and category values in the statement column.

Object	Object sentence	Slope @ I	Slope @ II	Slope @ III	Slope @ IV	Slope @ V
							1	AGMUNK	655	565	27	485	35
2	SR RBBB	413	278	55	322	199
							3	SR AMIAD	521	377	77	480	0
4	AGMUNK	0	567	1530	0	191
							5	AGMUNK	834	1211	950	594	0
Variance (variance)		78481	105809.4	372043.8	42902.6	8236.4

Table 2: representation of a reference data set comprising a plurality of features represented in columns and associated descriptive statistics Example (b)

Table 2 shows the data of table 1 with the variance of each column added in the last row of the table as descriptive statistical data for each characteristic ECG feature. Any suitable descriptive statistical data may be substituted for the variance. In addition, the category values in the form of statements in the statement column may be used to generate descriptive statistics, such as the occurrence of a given statement. For example, in table 2, 60% of the subjects appear with the statement AGMUNK.

In step 230, one or more first thresholds are defined, and in step 240, the variance of each of the plurality of normalized ECG data subsets is compared to one of the one or more first thresholds. The first threshold may be defined in a variety of ways depending on the application of the search feature extraction algorithm.

For example, defining one or more first thresholds may comprise defining a first global threshold for a plurality of normalized ECG data subsets, and wherein a variance of each ECG data subset of the plurality of ECG data subsets relating to the same characteristic ECG features as the plurality of normalized ECG data subsets is compared to the first global threshold.

Alternatively, defining the one or more first thresholds may comprise defining a first local threshold for each of the plurality of subsets of ECG data, respectively, and wherein the variance of each of the plurality of subsets of ECG data is compared to the first threshold defined for that subset of ECG data, respectively.

In step 250, at least one feature is extracted as a search feature based on the comparison.

In other words, extracting the search feature from the reference ECG data set comprises: determining a variance of the subset of ECG data; and if the variance is less than or equal to a predetermined threshold, identifying the subset as characteristic ECG features for extracting search features. However, if the variance is above a predetermined threshold, the subset may be rejected.

In other words, if the subset has a particularly high variance (i.e., a high degree of dissimilarity between the data within the subset), the data subset may not have useful features for searching and therefore will be rejected.

Alternatively or additionally, determining the characteristics of the subset of data may comprise: patterns (e.g., portions of an ECG waveform) in the subset of data are identified, and features of the patterns to be used as search features are determined.

In practice, the above-described search method and search feature extraction algorithm may be performed as follows. By way of example, fig. 3 shows a schematic view of a user interface 260 of a working example implementing the method described above.

The user interface shows a data table of data corresponding to a reference ECG data set, wherein the data table comprises data divided into columns corresponding to a plurality of features (feature 1, feature 2, etc.).

From a user interface, which may be implemented as a screen, touch screen, etc., a user may select a set of ECG data from the ECG data record to serve as the reference ECG data set. The set of data may be output from a previous search or may be imported directly. The set of data may have one or more unique ECG features of similar characteristics for further investigation. For example, data from five subjects contained the same diagnostic results.

After the user has selected the data as a reference ECG data set, the set of data is analyzed to obtain search criteria.

A search feature extraction algorithm is then applied to the reference ECG data set in order to extract the search features. This may be performed by first normalizing the values of each data type and calculating the variance of each normalized data type. Then, by means of a predefined or manually set threshold, a certain subset may be selected based on the calculated variance.

In the example shown in fig. 3, the user interface 360 includes a table for displaying the variance (Var) of each column of data, thereby displaying the variance value of each characteristic ECG feature of the reference ECG data set. In addition, the user interface displays an automatic recommendation rule for recommending a subset of data to be selected by the user based on the calculated variance. In the example shown in fig. 3, the recommended user selects a feature having a variance with an absolute value less than 0.2.

After the data subsets are selected, a series of values may be calculated for each data subset, which series of values can represent a characteristic of the data type, e.g., an average of the data types. This feature can then be used as a search feature to form the basis of a search criteria for searching the entire ECG database. For example, the criteria may be (absolute value (search value-feature 1) < threshold) and (absolute value (search value-feature 2) < threshold). After the search, the results may be searched for data similar to the reference ECG data set.

In the example shown in fig. 3, feature 1 and feature 2 are both selected by the user or automatically, e.g., because the variance of both features is less than 0.2. For example, features may be added or removed from a selected item by a user selecting an add or delete features button. The selected characteristic ECG features may then form part of a conditional formula for searching for similar data from the ECG database. As shown in fig. 3, the conditional formula specifies: for feature 1, the absolute value of the mean of the data obtained from the search in the ECG database must be less than 20 (feature 1ABS (mean-reference) < 20); whereas for feature 2, the absolute value of the mean of the data obtained from the search in the ECG database must be less than 30 (feature 2ABS (mean-reference) < 30). The user can customize the parameters of the condition accordingly. In addition, multiple selected subsets of data will cause the conditional formula to map to multiple conditions, and the logical relationship of these conditions may also be adjusted by the user or automatically.

After the conditions are set, the conditional function may be used as a search criteria, and the user may initiate a search of the ECG database, for example, by selecting a search database button. The internal authority can then search the ECG database for data that meets the conditions of the search criteria. After the search, the obtained relevant data may be displayed in a table on the user interface.

Alternative methods of identifying key features of the reference ECG data set may utilize one or more of the following: similar ECG features; abnormalities in the ECG data; and a diagnosis associated with the ECG data in the set. For example, if the ECG data in the reference ECG data set has a similar statement or diagnosis, a characteristic ECG feature related to such statement/diagnosis may be automatically recommended as the search feature. For example, in the case where all ECG data in the reference ECG data set show a myocardial infarction as a result, the ST segment of the ECG wave may be recommended as a search feature for a similar search.

Additionally, the ECG data may have similar features/patterns. For example, all ECG data in the reference ECG data set may have an inverse T wave or QRS notch. These features/patterns are then recommended for use in a similar search. Additionally, ECG data with similar abnormalities (e.g., high atrial rate (e.g., above 300/s and ST segment >0.2mV in V1)) may recommend atrial rate and ST segment for similar searches.

In a further alternative, the search feature may be manually selected. From a clinical perspective, certain data subsets in the reference ECG data set may have additional clinical significance or be clinically similar. For example, for a set of data including patients with the same symptoms or the same disease, the user may directly select pieces of data and use the pieces of data as search features. Features can then be computed according to the user's selection and used to search the ECG database.

More specifically, the user may directly select a certain number of columns or data subsets and use these columns as search features. After the columns are selected, a series of values may be calculated for each column that can represent a characteristic of the column, such as an average value for each column. The features of the selected column can then be used as search features to search the ECG database. For example, the search criteria may include (absolute value (measured value-feature 1 of all data) < threshold) and (absolute value (measured value-feature 2 of all data) < threshold).

A user interface may be provided to a user to customize the manner in which the calculations and/or the search criteria are established using the features of the reference ECG data set.

For example, the user may customize one or more thresholds as criteria for each search feature. The threshold may be customized for all selections made by the user, or if a group of data is selected by the user, the set of values corresponding to different sets of statistics may be customized.

In addition, the user interface can include one or more check boxes with which a research topic can be checked and, upon checking a research topic, measurements related to the research topic can be automatically selected. The user may also define one or more conditional relationships (e.g., and or other logical calculator) between features to form the final combined search criteria.

FIG. 4 illustrates an additional example 300 of a search feature extraction algorithm in accordance with an aspect of the subject invention.

In the example 300 shown in FIG. 4, the search feature extraction algorithm begins at step 310, where a first input indicative of a research topic is received from an interface at step 310.

For example, the user may be presented with a list of possible research topics available for selection or an interface that defines a customized research project.

In step 320, a plurality of characteristic ECG features associated with the study topic and a predefined rule for each of the plurality of characteristic ECG features are displayed at an interface. Predefined rules define known or expected relationships between pieces of data. These relationships may be based on clinical knowledge, physical laws, and the like.

In step 330, a second input indicative of the selected characteristic ECG feature and the set of rules is received. In other words, the user may select rules or relationships that are of interest to a particular research topic.

In step 340, the selected rule is applied to ECG data subsets within the reference ECG data set, wherein each ECG data subset is associated with one of the selected characteristic ECG features, and in step 350, when the percentage of the ECG data subsets associated with the one or more selected features complies with the selected rule, the one or more selected characteristic ECG features are extracted as search features.

Fig. 5 shows a schematic diagram of a user interface 360 implementing a working example of the method described above with reference to fig. 4.

As shown above in fig. 3, the user interface 360 shows a data table of data corresponding to a reference ECG data set, wherein the data table includes data divided into columns corresponding to a plurality of features (feature 1, feature 2, etc.). In addition, the user interface 360 includes a table for displaying the variance (Var) of each column of data, thereby displaying the variance value of each feature of the reference ECG data set.

In addition, the user interface displays automatic recommendation rules for recommending features to be selected by the user based on the indicated research topic. In the example shown in fig. 3, the user is recommended to select a feature related to Left Ventricular Hypertrophy (LVH).

In the example shown in fig. 3, feature 1 and feature 2 have both been selected by the user or automatically, for example because both features are related to LVH. For example, features may be added or removed from selected content by a user selecting an add or delete features button. The selected features may then form part of a conditional formula for searching for similar data from the ECG database. As shown in fig. 5, the conditional formula specifies: for feature 1, the absolute value of the mean of the data obtained from the search in the ECG database must be less than 20 (feature 1ABS (mean-reference) < 20); whereas for feature 2, the absolute value of the mean of the data obtained from the search in the ECG database must be less than 30 (feature 2ABS (mean-reference) < 30). The user can customize the parameters of the condition accordingly. In addition, multiple selected subsets of data will cause the conditional formula to map to multiple conditions, and the logical relationship of these conditions may also be adjusted by the user or automatically.

After the conditions are set, the conditional function may be used as a search criteria, and the user may initiate a search of the ECG database, for example, by selecting a search database button. The internal institution may then search the ECG database for data that meets the conditions of the search criteria. After the search, the obtained relevant data may be displayed in a table on the user interface.

FIG. 6 illustrates an example 400 of a search feature extraction algorithm in accordance with an aspect of the present invention.

In the example 400 shown in fig. 6, the search feature extraction algorithm begins at step 410 by obtaining a plurality of rules associated with a plurality of known characteristic ECG features, wherein each rule of the plurality of rules comprises a plurality of known clinical relationships, at step 410. The plurality of rules may include one or more of: diagnosing; a sentence; and a series of values.

In other words, a rule list may be obtained that includes a plurality of rules that may be generated based on clinical knowledge, research papers, clinical guidelines, and the like. For example, the rules may include: diagnostics, which may be indicated by statements such as LVH; ECG features as abnormal or rare cases according to clinical knowledge, e.g., tamp @ v1<0 (which corresponds to the amplitude of the T wave at a given time) or heart rate >150; or a combination of several criteria with known relationships, e.g. ston @ V1>2000 (where ston denotes the amplitude of the J-point of the ECG wave, which is the junction between the QRS complex and the ST segment) and tamp @ V1< -100.

In step 420, a plurality of rules are applied to a plurality of ECG data subsets within the reference ECG data set, wherein each of the ECG data subsets is associated with one of a plurality of known characteristic ECG features.

In step 430, one or more known characteristic ECG features are extracted as search features when the percentage of the ECG data subset associated with the known characteristic ECG features complies with the rules. For example, if 80% or more of the ECG data subsets comply with the rules, a characteristic ECG feature associated with the data subset may be selected as the search feature. The percentage threshold for extracting characteristic ECG features as search features may be adjusted by the user or automatically.

Table 3: example of a list of rules that can be applied to a reference ECG data set

In the example shown in Table 3, the Sokolow-Lyon standard is a set of standards for diagnosing LVH, where S V represents the S-wave amplitude at lead V1, and R V represents the R-wave amplitude at lead V5. The Cornell Voltage criteria is an alternative set of criteria for diagnosing LVH, where R aVL represents the R-wave amplitude at lead aVL.

Based on different criteria, the ECG system may have different statements related to LVH, as shown by the statements in the LVH diagnostic line above. These statements are given by the LVH related algorithm.

The above methods may be used in combination with each other or independently.

Fig. 7 shows an example of a computer 500 for implementing the above-described method. The computer may be part of a cloud computing environment, a server, or a standalone terminal.

Computer 500 includes, but is not limited to, a PC, workstation, laptop, PDA, palm device, server, storage device, and the like. Generally, in terms of hardware architecture, computer 500 may include one or more processors 501, memory 502, and one or more I/O devices 503 communicatively coupled via a local interface (not shown). The local interface can be, but is not limited to, one or more buses or other wired or wireless connections as is known in the art. The local interface may have additional elements (e.g., controllers, buffers (caches), drivers, repeaters, and receivers) to enable communications. In addition, the local interface may include address, control, and/or data connections to enable appropriate communications among the aforementioned components.

The processor 501 is a hardware device for running software that can be stored in the memory 502. The processor 501 can be almost any custom made or commercially available processor, central Processing Unit (CPU), digital Signal Processor (DSP) or an auxiliary processor among several processors associated with the computer 500, and the processor 501 may be a semiconductor based microprocessor (in the form of a microchip) or a microprocessor.

The memory 502 can include any one or combination of the following: volatile memory elements (e.g., random Access Memory (RAM), such as Dynamic Random Access Memory (DRAM), static Random Access Memory (SRAM), etc.) and nonvolatile memory elements (e.g., ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), tape, compact disc read-only memory (CD-ROM), magnetic disk, floppy disk, cartridge, tape cartridge, etc.). Moreover, the memory 502 may include electronic, magnetic, optical, and/or other types of storage media. Note that the memory 502 can have a distributed architecture, where various components are located remotely from each other, but can be accessed by the processor 501.

The software in memory 502 may include one or more separate programs, each of which comprises an ordered listing of executable instructions for implementing logical functions. The software in memory 502 includes a suitable operating system (O/S) 504, compiler 505, source code 506, and one or more applications 507 in accordance with the illustrative embodiments.

The application 507 includes a number of functional components, such as computing units, logic units, functional units, processes, operations, virtual entities, and/or modules.

Operating system 504 controls the execution of computer programs and provides scheduling, input-output control, file and data management, memory management, and communication control and related services.

Application 507 may be a source program, executable program (object code), script, or any other entity comprising a set of instructions to be executed. When a source program, the program is typically translated via a compiler (e.g., compiler 505), assembler, interpreter, or the like, which can be included within memory 502 or non-included within memory 502, such that the program operates properly in connection with operating system 504. Further, the application 507 can be written as an object oriented programming language (which has categories of data and methods) or a procedural programming language (which has routines, subroutines, and/or functions, such as, but not limited to C, C + +, C #, pascal, BASIC, API calls, HTML, XHTML, XML, ASP script, javaScript, FORTRAN, COBOL, perl, java, ADA,. NET, etc.).

The I/O devices 503 may include input devices such as, but not limited to, a mouse, a keyboard, a scanner, a microphone, a camera, and the like. Further, I/O devices 503 may also include output devices such as, but not limited to, a printer, a display, and the like. Finally, I/O devices 503 may also include devices that communicate input and output, such as, but not limited to, network Interface Controllers (NICs) or modulators/demodulators (for accessing remote devices, other files, devices, systems, or networks), radio Frequency (RF) or other transceivers, telephony interfaces, bridges, routers, and the like. The I/O device 503 also includes components for communicating over various networks, such as the internet or an intranet.

When the computer 500 is running, the processor 501 is configured to run software stored within the memory 502 to communicate data bi-directionally with the memory 502 and to control the operation of the computer 500 generally in accordance with the software. Applications 507 and operating system 504 are read in whole or in part by processor 501, possibly cached within processor 501, and then executed.

When the application 507 is implemented in software, it should be noted that the application 507 can be stored on virtually any computer-readable medium for use by or in connection with any computer-related system or method. In the context of this document, a computer readable medium may be an electronic, magnetic, optical, or other physical device or means that can contain or store a computer program for use by or in connection with a computer related system or method.

Variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality.

A single processor or other unit may fulfill the functions of several items recited in the claims.

The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the internet or other wired or wireless telecommunication systems.

If the term "adapted" is used in the claims or the description, it should be noted that the term "adapted" is intended to be equivalent to the term "configured to".

Any reference signs in the claims shall not be construed as limiting the scope.

Claims

1. A computer-implemented method (100) for searching a database of ECG data, the method comprising:

obtaining (110) a reference ECG data set, wherein the reference ECG data set comprises ECG data of a plurality of subjects, the ECG data of the plurality of subjects comprising data values of a plurality of characteristic ECG features of each of the subjects;

grouping (115) the reference ECG data set into a plurality of ECG data subsets, wherein each of the plurality of ECG data subsets comprises data values of a respective one of a plurality of characteristic ECG features;

applying (120) a search feature extraction algorithm to the reference ECG data set to extract at least one of the characteristic ECG features as a search feature, wherein a search feature is a characteristic ECG feature that enables identification of ECG data of interest within the database of ECG data;

generating (130) search criteria based on the search features; and is

Searching (140) the database using the search criteria to obtain ECG data of interest, wherein the search feature extraction algorithm comprises:

normalizing (210) each of the plurality of ECG data subsets to generate a plurality of normalized ECG data subsets;

calculating (220) a variance of each normalized ECG data subset of the plurality of normalized ECG data subsets;

defining (230) one or more first thresholds;

comparing (240) the variance of each of the plurality of normalized ECG data subsets to one of the one or more first thresholds; and is

At least one feature is extracted (250) as a search feature based on the comparison.

2. The computer-implemented method (100) of claim 1, wherein defining one or more first thresholds comprises defining a first global threshold for a plurality of normalized ECG data subsets, and wherein the variance of each of the plurality of ECG data subsets is respectively compared to the first global threshold.

3. The computer-implemented method (100) of claim 1, wherein defining one or more first thresholds comprises defining a first local threshold for each of the plurality of subsets of ECG data, and wherein the variance of each of the plurality of subsets of ECG data is compared to the first local threshold defined for that subset of ECG data.

4. The computer-implemented method (100) of claim 1, wherein normalizing each of the plurality of subsets of ECG data comprises:

normalizing the subsets of ECG data based on maximum and minimum values of the database of ECG data associated with the same characteristic ECG features as the plurality of subsets of ECG data; or

5. The computer-implemented method (100) of any of claims 1 to 4, wherein generating the search criteria comprises:

defining a second threshold for the search feature; and is

Generating the search criteria based on the average and the second threshold.

6. The computer-implemented method (100) of claim 1, wherein the reference ECG data set further comprises ECG data of a plurality of subjects, the ECG data of the plurality of subjects comprising category values of a plurality of characteristic ECG features of each of the subjects, and the search feature extraction algorithm comprises:

receiving (310) a first input from an interface indicative of a research topic;

displaying (320) a plurality of characteristic ECG features associated with the study subject and a predefined rule for each of the plurality of characteristic ECG features at the interface;

receiving (330) a second input indicative of the selected characteristic ECG features and the set of rules;

applying (340) the selected rule to the ECG data subsets within the reference ECG data set, wherein each ECG data subset is associated with one of the selected characteristic ECG features; and is provided with

Extracting (350) one or more selected characteristic ECG features as search features when a percentage of the ECG data subset associated with the one or more selected characteristic ECG features complies with the selected rule.

7. The computer-implemented method (100) of claim 1, wherein the reference ECG data set further comprises ECG data of a plurality of subjects, the ECG data of the plurality of subjects comprising category values of a plurality of characteristic ECG features of each of the subjects, and the search feature extraction algorithm comprises:

obtaining (410) a plurality of rules associated with a plurality of known characteristic ECG features;

applying (420) the plurality of rules to at least one of the plurality of ECG data subsets within the reference ECG data set, wherein each of the at least one of a plurality of ECG data subsets is associated with a characteristic ECG feature of the plurality of known characteristic ECG features; and is

Extracting (430) one or more known characteristic ECG features as search features when the percentage of the ECG data subset associated with the known characteristic ECG features complies with the rules, wherein the plurality of rules include one or more of:

diagnosing;

a sentence; and

a range of values.

8. The computer-implemented method (100) according to any one of claims 1 to 7, wherein the method further comprises:

presenting a plurality of search features to a user; and is

9. A computer program comprising computer program code means adapted to perform the method of any of claims 1 to 8 when said computer program is run on a computer.

10. A system for searching a database of ECG data, the system comprising:

a processor adapted to:

generating search criteria based on the search features; and is provided with

Searching the database using the search criteria to obtain ECG data of interest,

wherein, when applying the search feature extraction algorithm, the processor is adapted to:

calculating a variance of each of the plurality of normalized ECG data subsets;

defining one or more first thresholds;

At least one feature is extracted as a search feature based on the comparison.

11. The system of claim 10, wherein the reference ECG data set further comprises ECG data of a plurality of subjects including category values of a plurality of characteristic ECG features, and the processor, when applying the search feature extraction algorithm, is adapted to:

receiving a first input from an interface indicative of a research topic;

applying the selected rule to the ECG data subsets within the reference ECG data set, wherein each ECG data subset is associated with one of the selected characteristic ECG features; and is provided with

12. The system of claim 10, wherein the reference ECG data set further comprises ECG data of a plurality of subjects, the ECG data of the plurality of subjects comprising category values of a plurality of characteristic ECG features of each of the subjects, and the processor is adapted to, when applying the search feature extraction algorithm:

obtaining a plurality of rules associated with a plurality of known characteristic ECG features;

Extracting one or more known characteristic ECG features as search features when the percentage of the ECG data subset associated with the known characteristic ECG features complies with the rules,

wherein the plurality of rules comprise one or more of:

diagnosing;

a sentence; and

a range of values.